When Puffing Meets Strombolian Explosions: A Tale of Precursors and Coda

Abstract:

Normal activity at Stromboli (Aeolian Islands, Italy) can be broadly divided into (1) puffing, where small amounts of gas are emitted without pyroclasts at a rate of ~1 per second; and (2) "normal" explosions, where gas and pyroclast mixtures are expelled up to several hundred meters high (at rates of 1-10 events per hour).

In May 2014, we monitored Stromboli's active vents with high speed visible-light and thermal infrared cameras. At vents in the SW, central and NE craters, puffing was clearly visible, emitting hot gas plumes lasting less than 0.2 s and sometimes spaced by less than 0.1 s, causing frequent event overlap. The initial ascent velocity of the gas was ~10 m/s, typical of a jet phase produced by burst of a slightly overpressurized bubble. The variability of the emission point within a vent suggested that these bubbles were smaller than the vent radius. The size of individual puffs estimated from the radiant energy of the plume was variable, but limited by an upper bound depending on the vent characteristics.

Meanwhile, "normal" explosions occurred at intervals of a few minutes. The high ejection velocity of the pyroclasts (up to 400 m/s) and the duration of the explosions (5-30 s) suggested that explosions came from large and highly pressurized slugs arriving at the "free surface".

Explosion onsets were well defined by the emission of the first, fast, pyroclasts. However, in most cases, precursory gas was recorded up to 10 s before pyroclast release. The absence of pyroclasts and the initial ascent velocity of gas indicated that these emissions corresponded to puffing-like activity. On the other hand, their size were greater than regular puffs (sometimes by more than one order of magnitude) and frequently increased with time. This suggested that these precursory puffs were strongly influenced by the rising slug about to burst.

Similarly, after the last pyroclast ejection, large and low-velocity gas release events were occasionally observed, and interpreted as a coda due to the magma disturbance generated by the explosion. Precursors and coda can be used to estimate the changes in the pressure field produced by the rise and burst of slugs.